# Save a multidimensional array with metadata to data in NetCDF format This function allows to save a data array with metadata into a NetCDF file, allowing to reload the saved data using `Start` function from StartR package. If the original 's2dv_cube' object has been created from `CST_Load()`, then it can be reloaded with `Load()`. ```r SaveExp( data, destination = tempdir(), coords = NULL, Dates = NULL, time_bounds = NULL, startdates = NULL, varname = NULL, metadata = NULL, Datasets = NULL, sdate_dim = "sdate", ftime_dim = "time", memb_dim = "member", dat_dim = "dataset", var_dim = "var", drop_dims = NULL, single_file = FALSE, extra_string = NULL, global_attrs = NULL, units_hours_since = FALSE ) ``` ## Arguments - `data`: A multi-dimensional array with named dimensions. - `destination`: A character string indicating the path where to store the NetCDF files. - `coords`: A named list with elements of the coordinates corresponding to the dimensions of the data parameter. The names and length of each element must correspond to the names of the dimensions. If any coordinate is not provided, it is set as an index vector with the values from 1 to the length of the corresponding dimension. - `Dates`: A named array of dates with the corresponding sdate and forecast time dimension. If there is no sdate_dim, you can set it to NULL. It must have ftime_dim dimension. - `time_bounds`: (Optional) A list of two arrays of dates containing the lower (first array) and the upper (second array) time bounds corresponding to Dates. Each array must have the same dimensions as Dates. If 'Dates' parameter is NULL, 'time_bounds' are not used. It is NULL by default. - `startdates`: A vector of dates that will be used for the filenames when saving the data in multiple files (single_file = FALSE). It must be a vector of the same length as the start date dimension of data. It must be a vector of class `Dates`, `'POSIXct'` or character with lenghts between 1 and 10. If it is NULL, the coordinate corresponding the the start date dimension or the first Date of each time step will be used as the name of the files. It is NULL by default. - `varname`: A character string indicating the name of the variable to be saved. - `metadata`: A named list where each element is a variable containing the corresponding information. The information must be contained in a list of lists for each variable. - `Datasets`: A vector of character string indicating the names of the datasets. - `sdate_dim`: A character string indicating the name of the start date dimension. By default, it is set to 'sdate'. It can be NULL if there is no start date dimension. - `ftime_dim`: A character string indicating the name of the forecast time dimension. By default, it is set to 'time'. It can be NULL if there is no forecast time dimension. - `memb_dim`: A character string indicating the name of the member dimension. By default, it is set to 'member'. It can be NULL if there is no member dimension. - `dat_dim`: A character string indicating the name of dataset dimension. By default, it is set to 'dataset'. It can be NULL if there is no dataset dimension. - `var_dim`: A character string indicating the name of variable dimension. By default, it is set to 'var'. It can be NULL if there is no variable dimension. - `drop_dims`: (optional) A vector of character strings indicating the dimension names of length 1 that need to be dropped in order that they don't appear in the netCDF file. Only is allowed to drop dimensions that are not used in the computation. The dimensions used in the computation are the ones specified in: sdate_dim, ftime_dim, dat_dim, var_dim and memb_dim. It is NULL by default. - `single_file`: A logical value indicating if all object is saved in a single file (TRUE) or in multiple files (FALSE). When it is FALSE, the array is separated for datasets, variable and start date. When there are no specified time dimensions, the data will be saved in a single file by default. The output file name when 'single_file' is TRUE is a character string containing: '__.nc'; when it is FALSE, it is '_.nc'. It is FALSE by default. - `extra_string`: (Optional) A character string to be included as part of the file name, for instance, to identify member or realization. When single_file is TRUE, the 'extra_string' will substitute all the default file name; when single_file is FALSE, the 'extra_string' will be added in the file name as: '__.nc'. It is NULL by default. - `global_attrs`: (Optional) A list with elements containing the global attributes to be saved in the NetCDF. - `units_hours_since`: (Optional) A logical value only available for the case: Dates have forecast time and start date dimension, single_file is TRUE and 'time_bounds' is NULL. When it is TRUE, it saves the forecast time with units of 'hours since'; if it is FALSE, the time units will be a number of time steps with its corresponding frequency (e.g. n days, n months or n hours). It is FALSE by default. ## Returns Multiple or single NetCDF files containing the data array. - **`single_file is TRUE`**: All data is saved in a single file located in the specified destination path with the following name (by default): '__.nc'. Multiple variables are saved separately in the same file. The forecast time units are calculated from each start date (if sdate_dim is not NULL) or from the time step. If 'units_hours_since' is TRUE, the forecast time units will be 'hours since '. If 'units_hours_since' is FALSE, the forecast time units are extracted from the frequency of the time steps (hours, days, months); if no frequency is found, the units will be ’hours since’. When the time units are 'hours since' the time ateps are assumed to be equally spaced. - **`single_file is FALSE`**: The data array is subset and stored into multiple files. Each file contains the data subset for each start date, variable and dataset. Files with different variables and datasets are stored in separated directories within the following directory tree: 'destination/Dataset/variable/'. The name of each file will be by default: '_.nc'. The forecast time units are calculated from each start date (if sdate_dim is not NULL) or from the time step. The forecast time units will be 'hours since '. ## Examples ```r data <- lonlat_temp_st$exp$data lon <- lonlat_temp_st$exp$coords$lon lat <- lonlat_temp_st$exp$coords$lat coords <- list(lon = lon, lat = lat) Datasets <- lonlat_temp_st$exp$attrs$Datasets varname <- 'tas' Dates <- lonlat_temp_st$exp$attrs$Dates metadata <- lonlat_temp_st$exp$attrs$Variable$metadata SaveExp(data = data, coords = coords, Datasets = Datasets, varname = varname, Dates = Dates, metadata = metadata, single_file = TRUE, ftime_dim = 'ftime', var_dim = 'var', dat_dim = 'dataset') ``` ## Author(s) Perez-Zanon Nuria, nuria.perez@bsc.es

SaveExp function